A major limitation to the widespread use of action potential-prolonging drugs is the occasional, and apparently unpredictable, development of the polymorphic ventricular tachycardia Torsades de Pointes during therapy. Anecdotes have suggested that patients with atrial fibrillation are relatively "immune" to this adverse drug effect, while those in whom atrial fibrillation has been converted to normal rhythm are at particularly high risk in the immediate post-conversion period. To test this hypothesis, subjects in whom atrial fibrillation has been converted to normal rhythm are at particularly high risk in the immediate post-conversion period. To test this hypothesis, subjects with atrial fibrillation scheduled for DC cardioversion are receiving intravenous infusions of the action potential-prolonging agent dofetilide on two occasions: once during trial fibrillation just prior to cardioversion and the second time during sinus rhythm within 24 hours of cardioversion. To date, 12 subjects have participated in this study. Eight subjects have received both infusions, while in 4 sinus rhythm could not be restored by either drug infusion or DC cardioversion. In these eight, dofetilide did not prolong the QT interval when administered during atrial fibrillation, but markedly prolonged the QT interval when the underlying rhythm was sinus. One subject had an asymptomatic 12-second episode of Torsades de Pointes during the sinus rhythm infusion. Thus, this study has demonstrated that patients with atrial fibrillation are, indeed, more sensitive to the action potential prolonging effects of antiarrhythmic drugs in the immediate post-cardioversion period. The mechanism underlying this effect will require analyses of serum dofetilide, catecholamine, naturetic peptide, and electrolyte concentrations, which are underway.